The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data, specifically Job Descriptions (JD) and user messages.
Ingestion points: Untrusted data enters the agent context through the userMessage and jd input fields defined in schema.json.
Boundary markers: The instructions do not specify explicit boundary markers or delimiters to isolate untrusted input.
Capability inventory: The skill has no execution capabilities; it does not utilize shell commands, file writing, or network operations.
Sanitization: There is no explicit sanitization step described, although the output is constrained to a rigid JSON schema.
Assessment: Because the skill lacks the capability to act on instructions (no tool access), the risk of indirect prompt injection is negligible.
[SAFE]: The skill instructions and schema do not contain any hardcoded credentials, obfuscated content, or persistence mechanisms. All logic is focused on natural language processing and structured data extraction for career consulting.

dig